The work explores the three-dimensional unsteady wake of a flapping-wing Micro Air Vehicle (MAV) ‘DelFly II’, applying a Kriging regression technique for the spatial regression of time-resolved Stereoscopic Particle Image Velocimetry (Stereo-PIV) data. In the view of limited number of measurement planes, the particular objective of the regression is to provide an accurate volumetric representation of the measurement domain on a spatial grid that is much finer than the spacing between the measurement planes. A unique feature of the current study is the incorporation of a statistical error model in the Kriging regression process as an estimate of the local measurement uncertainty of the PIV measurements. The performance of the Kriging regression technique with local error estimates is evaluated based on direct comparisons to measurement data. As a final result of the regression, three-dimensional vortical structures are reconstructed and it is shown that each of two wings generates separate vortex loops and sheds two trailing edge vortices during the downstroke phase. On the other hand, the structures of upper and lower wings occasionally interact with each other during the upstroke phase. Wake structure displays major differences in terms of vortex formations for different reduced frequencies.